YOU ARE AT:Telco CloudHybrid vs. public—How should operators move to the telco cloud? 

Hybrid vs. public—How should operators move to the telco cloud? 

As operators move into the telco cloud era, which deployment option(s) will win out? 

As the title of this piece—and the Telco Cloud & Edge Forum session it’s based on—suggest, it seems that the two most likely outcomes for operators on a long-term journey to cloud-native are to run IT and network workloads in a public cloud, or in a combination of public and private clouds with a unified management layer. The option that’s missing is an all-private approach to the telco cloud. While this is a departure from the real world where real operators run all their workloads in fortified, customized private clouds, we convened two leading operators with one of the biggest cloud players in the game and there was agreement that public and hybrid beat private cloud every time. 

The contours of this debate where aptly set up by STL Partners Senior Analyst David Martin. On the one hand, private is private; on the other hand, public clouds provided by hyperscalers open up lots of opportunity around analytics, artificial intelligence (AI), resiliency and other indisputable positives. What about combining the best of both into a hybrid cloud? “Is a hybrid approach actually the most effective strategy, a way to promote agility, flexibility and resilience while mitigating the risks of over-reliance on a single cloud provider? Or is it rather the product of a compromise and a cherry-picking attitude which telcos have been blame for in other areas in the past? Which means that telcos may fail to fully realize the potential of cloud to drive efficiencies of scale and consolidation of workloads on a unified platform.” 

Pradeep Jhunjhunwala, vice president and head of cloud engineering at Reliance Jio provided the telco cloud perspective of one of the world’s largest operators, which also has a rich history of building and operating networks using cutting-edge architectures and processes, and which also has a huge opportunity to provide advanced network-based services to a vibrant enterprise base. From his point of view, “This is not either/or; this is and…You can’t depend exclusively on public cloud and you can’t depend completely on hybrid cloud. You need to have a mix of both private and public cloud and build a hybrid cloud.” 

Jhunjhunwala also shifted from the operator the enterprise user point of view. “The OT workload can’t run on public cloud because we see a latency requirement…but you have [an] IT workload which can run on [public] cloud. So this is a world which is going to survive in hybrid cloud. [It] can’t be just a public cloud world because of multiple reasons…either this is regulation issues or latency issues or performance issues…Based on the user demand for latency, performance, cost regulation, all these different models evolve. And what enterprise has to do, or any customer has to do, is basically look at those scenarios and adopt the right cloud framework.” 

For Telenor, Director of Cloud Strategy and Architecture Pål Grønsund said the operator has elected to follow a “public cloud-first strategy, basically because there are so many benefits like the scalability, flexibility, automation and other types of services that we consume while running there.” That said, Telenor does have network workloads in its private cloud, although that is gradually changing; he noted Telenor runs its MVNO’s 5G core in AWS’s public cloud, for example. Telenor is also testing multi-vendor 5G core in private and hybrid telco cloud environments—all in all, “It’s really moving forward.” 

Progress, however, is not without challenges, Grønsund said, specifically honing in on regulatory-related obstacles. He called out “national autonomy,” as in some externality like a war cutting off access to a core network running in the public cloud. Or, in another example, the need for “local key encryption systems” covered in a contractual arrangement. Another one: “There’s the need for a certain level of security-cleared personnel operating the cloud even onto the physical hardware, something we need to guarantee.”

Is the real question around the future, or lack thereof, of an all-private approach telco cloud? 

Before digging into the question above, Microsoft’s Shawn Hakl, vice president of product management, reframed hybrid or public as physics and security. “It’s going to be hybrid for the foreseeable future, probably forever, right? You’re not going to get the physics to change between say, for example, the radio unit and the [distributed unit].” 

He highlighted Microsoft’s work with AT&T to move many of the operator’s workloads, both network and IT, into the Azure cloud. As an aside, AT&T did extensive R&D and commercialization of its own telco cloud platform which it sold to Microsoft. He said AT&T’s Standalone 5G workloads, for instance, are in a hybrid cloud. “We have deployments of technology that are Microsoft managed within AT&T facilities. So that gives us geographic diversity and that’s all managed from an Azure public cloud instance. And so then they have the option to move functionality into the public cloud depending on what makes sense for them as a business. What it does do is it leverages public cloud services. So I think an important part here is the difference between when people talking about private cloud versus hybrid cloud. I think the real debate is are people going to continue to build their own private clouds?” 

Why would the people stop building private clouds? More AI-poweed tools that map to a reduction of opex, as well as other analytic and zero-trust security functionality, Hakl said. Operators are “going to leverage the services that come out of the cloud to do that. It doesn’t mean that they’re going to physically move all of their network functions off premise.” 

He noted Azure’s workflow of “constantly implementing changes” operators can leverage at rapidly and at scale. “I think the trend towards going form just merely getting technology deployed to having to scale the business, lowering the cost per bit, will drive people to adopt these technologies more often.” But, “You’ve got to give, essentially, the ability to seize control of those resources that are running their network workloads and bring them back under control temporarily and then potentially migrate them back to the hyperscale [environment in the event of a national emergency,” for example. “Those are the kind of lessons learned we’ve got working with ATYT and others as we deploy this technology at scale.” 

As summed up by Martin, there is “consensus emerging around the fact that it’s not really an either/or; it’s not either hybrid or public exclusively that we are working towards.” An “environment that is essentially hybrid but with a perhaps growing role for public crowd,” if you will.

Extending the telco cloud to the network edge

Jhunjhunwala sees the rise of connectivity and mobile edge computing for enterprises working on Industry 4.0-type use cases as a vector for hyperscalers to increasingly influence the traditional telecoms domain. Latency KPIs for OT workloads—robotic control in a manufacturing facility, for example—might require a new approach to not just connectivity but also location and control of cloud computing resources. 

“We’ll have regional edges coming up which is going to serve that,” he said. “And this can be hyperscaler edge or this can be a private edge, which the telcos are going to build because I firmly believe that telcos are the right people to build the edge technologies either within partnership with the hyperscalers, or they can build their own stack. Both will be required.” Jhunjhunwala predicted a shift from centralized to distributed public cloud resources in a five- to seven-year timeframe. 

The requirements of industries, he said, again reiterating latency requirements, will ultimately result in “cloud next to your requirement.” 

Hakl identified hyperscalers’ lack of spectrum as an indicator that operators will always have a part to play, which addresses an industry concern around whether the cloud giants are coming for operators’ customers. “We’re not connected to anybody that doesn’t come through an operator. And from our perspective…there’ll be a highly distributed fabric composed of compute and communications resources that span from 5G into space.” 

Bottomline on telco cloud, he said, “We think operators are natural partners for us…Operators have a unique view of intelligence that comes out of the network and a unique set of functionality…We fully expect every operator to be partnering with every cloud provider provider because, let’s face it, their customers already partner with all of us. We have an opportunity to grow the pie together. I think this is where the excitement comes in.”

Click here for more telco cloud news and fundamentals content.

ABOUT AUTHOR

Sean Kinney, Editor in Chief
Sean Kinney, Editor in Chief
Sean focuses on multiple subject areas including 5G, Open RAN, hybrid cloud, edge computing, and Industry 4.0. He also hosts Arden Media's podcast Will 5G Change the World? Prior to his work at RCR, Sean studied journalism and literature at the University of Mississippi then spent six years based in Key West, Florida, working as a reporter for the Miami Herald Media Company. He currently lives in Fayetteville, Arkansas.